Wolfram|Alpha is the newest brainchild of Steven Wolfram.  If you are unfamiliar with his work, he is the man responsible for Mathematica, the most common mathematical programming language to find its way into high schools and colleges.

Wolfram|Alpha, which is scheduled to launch in about a week, is a new kind of search engine designed to find and formulate answers to any sort of factual question.  In short, it’s supposed to act as a computerized subject area expert.

Always eager to build on his previous work, Steven Wolfram constructed Wolfram|Alpha inside Mathematica, utilizing some five to six million lines of code.  By way of comparison, that’s about fifty times as much as our Content Management System.

Anyway, last week I had the good fortune to watch a sneak preview of Wolfram Alpha (now conveniently available on You Tube).

Most of the demonstration focused on demonstrating the breadth of topics which Wolfram|Alpha can already find answers for, and the versatility (and limitations) of it’s natural language algorithm.

For those of you who have not made a study of Search Engine Technologies, natural language processing, or the ability to understand everyday speech, is one of the holy grails that all the major companies have been seeking.

However, impressive as that accomplishment might be to specialists, it doesn’t really address the questions that business professionals are likely to have.

1)    How will research with Wolfram|Alpha differ from research with a traditional search engine like Google?

Most traditional search engines will look for popular (and they hope well regarded) experts whose websites include text close to the search phrase put in by the user.  These pages may or may not actually answer the user’s question, or they may be incomplete, and require the use of other sites to fill in the context.

Wolfram|Alpha will take that same input, try to figure out what question or questions the user was trying to ask, and then use a number of professionally vetted sources to provide models, charts, formulae, and specific answers to the likely questions.  When there are major differences of opinion, Wolfram|Alpha is supposed to ask which model or set of assumptions you want it to use.

2)    Can I ask Wolfram|Alpha questions about my company or my competition?

Right now it will depend on how well documented the relevant companies are within the literature and feeds which Wolfram|Alpha is already tapping.

For example, Wolfram|Alpha can answer financial questions about companies with publicly listed stock, but it may not know how many restaurants sell baked ziti in the greater Bay Area.

3)    Can I tell Wolfram about my business so that it will show up when people ask relevant questions?

Steven Wolfram indicated that this option will not be available when Wolfram|Alpha launches, but that he does intend to eventually release a commercial analog.  When that happens, there will be a process whereby individual companies will be able to code in the relevant information so that the automated aspects of Wolfram|Alpha’s  processes will be able to prepare the data for use.

Because the natural language processing that Wolfram formulated is specifically designed to help people asking questions (rather than providing answers), this information will need to be coded, and will most likely become a part of the SEO Industry.

If you have any further questions about Wolfram|Alpha, ask them here, and I’ll see if I can find you an answer.